ABSTRACT
we have developed an algorithm based on Probabilistic Classification to generate Decision Trees. The core of the paper deals with the determination of the Most Significant Attribute done by segmenting the data obtained from a huge data warehouse. The attributes would be ranked based on the error rate and based on it the decision trees would be generated, which helps in elimination of zero occurrence templates.
There are chances for the occurrence of dual ranks for the attributes for which MSA is calculated. The duality is resolved by using fuzzy logic, which in turn optimizes the mining process. Traversing through the Decision tree, we can predict values for unknown attributes.
Our algorithm is meant to bridge the gap between such small Marketing organizations and the technology of Data Mining.
P.BHARANI,
S.ASRAF NAVEEN.
0 Comments